Can gender and race dynamics in performance appraisals be disrupted? The case of anchoring

Performance reviews in firms are common but controversial. Managers’ subjective appraisals of their employees’ performance and employees’ self-evaluations might be affected by demographic characteristics. As self-evaluations are typically shared with managers before they rate employees, social influence via anchoring may also contribute to gender or race differences in performance ratings. Analyzing the data of a multi-national financial services firm, we find gender and intersectionality gaps in self-ratings as well as race and intersectionality gaps in manager ratings, leading to female employees of color being assigned the lowest final ratings. We evaluate a mechanism that could disrupt these dynamics: a quasi-exogenous shock leading to self-evaluations not being shared with managers before they appraised employees. This intervention induced “de-anchoring,” leading to lower average manager ratings. However, it did not change overall gender or race dynamics as managers imposed their own gender and race effects (independent of self-evaluations), and also had access to previous years’ ratings for most employees. To exclude potential anchors from the past, we focus on employees with no history in the firm during their first year of employment. Among these newcomers, employees who traditionally have given themselves the lowest self-ratings—women of color—benefitted most from the intervention. The intervention did not, however, close the manager-induced race gap overall. These race dynamics were particularly pronounced in the US, negatively affecting Black, Asian and Latinx employees. Counterfactual simulations suggest that 22-28% of Black employees’ manager ratings would have to be increased for this race gap to disappear.